A key element of study of the biomechanics of pedestrian injuries from vehicle collisions is the determination of pre-impact vehicle speed. Pedestrian projection distances are an important means to determine collision speed, particularly as tyre brake marks are not readily observable with ABS brakes. A number of models have been developed by the authors, which include recent analytical models for forward [1] and wrap projection [2] impact. These models are novel as they include explicit modelling of the impact phase. They have been validated against the available test data, showing very good comparisons, and are therefore ideal for further statistical analysis. Confidence limits for speed estimates are set out for various purposes including injury research and litigation. The models show that the distribution of predicted collision speeds from projection distance can be large when a high degree of confidence is required. Some of this uncertainty is due to the impact phase where parameters such as duration and restitution are unknown for individual collisions, in addition to other confounding factors that are difficult to quantify. However the effects of the coefficient of retardation during the projection phase and the mass ratio between pedestrian and the striking vehicle can be readily determined. This paper analyses the influence of reduced variability of the input parameters on the predicted range of impact velocities. Analysis for known coefficients of retardation and mass ratio values yields the minimum prediction uncertainty due to the variability of the impact phase parameters alone. Results show that significant improvements in prediction uncertainty can be achieved by exact knowledge of the pedestrian to vehicle mass ratio and of the coefficient of retardation between the pedestrian and the road surface. In practice, while the mass ratio can be determined for individual collisions there is significant uncertainty as to the coefficient of retardation values, as these are influenced by the kinematics of the pedestrian during projection and ground impacts, in addition to factors such as road surface condition and contamination. However, overall data is available for wet and dry road conditions. Tables are presented for impact speed prediction from projection distance for various conditions and confidence levels. (Author/publisher)
Abstract